Ultrafiltration in the Branchial Heart Appendage of Dibranchiate Cephalopods: A Comparative Ultrastructural and Physiological Study

As in vertebrates, urine formation in the renal organs of various higher invertebrates is believed to be a consequence of three processes : ultrafiltration, active reabsorption and secretion. This assumption applies particularly to various groups of arthropods and molluscs (Rudell & Wellings, 1971 ; Riegel & Cook, 1975; Potts, 1975), especially to the highly organized cephalopods. The branchial heart appendages of the genus Sepia and Octopus have been shown to contain transporting cell types and podocytes, which are typical of the vertebrate nephron (Kümmel, 1967 ; Schipp, Hohn & Schafer, 1971 ; Witmer & Martin, 1973 ; Schipp & Boletzky, 1975); and ultrafiltration has been indicated by inulin-clearance in Octopus dofleini (Harrison & Martin, 1965). However, it is not clear how an ultrafiltrate can be formed since the inner surface of the appendage seems to be completely covered by the transporting epithelium (Schipp et al. 1971 ; Witmer & Martin, 1973).

In the present study a possible route for the filtrate is shown, and the hydrostatic pressure across the appendage wall is shown sometimes to exceed the osmotic pressure difference.

Fifty juvenile and adult animals of both sexes of the species Octopus vulgaris (Lam.) and Sepia officinalis (L.), and some adult Eledone cirrosa (Lam.) from the Bassin d’Arcachon (Atlantic Ocean) and from the Mediterranean near Banyuls-sur-Mer, were used in this study.

Length, weight and sex of the animals of the principal experiments are listed in Table 1.

Material fixed in buffered formalin or Bouin’s solution was embedded in paraffin or Araldite. Paraffin sections (10 μm) were stained with Masson’s trichrome modified by Goldner, or according to Bodian’s method or with alcian blue. Araldite sections ( 1 μm) were studied in the phase contrast microscope.

Tissue was prefixed in 4 % glutaraldehyde in phosphate buffer (Sorensen, modified; 1000 m-osmol/l; pH 7·2; 2 h; 6–8 °C) followed by a post-fixation in 1·5% OsO₄ in phosphate buffer for 1–2 h. The material was embedded in Araldite. Ultra-thin sections were cut on a LKB or Reichert ultramicrotome, stained with uranyl acetate and lead citrate and viewed in a Zeiss EM 9 A or a Philips EM 300.

1. Measurements of hydrostatic pressures were carried out with seven animals narcotized by 1 % ethanol, with opened mantle and under rinsing of the gills with gassed sea water. The pressures were measured directly by puncture of the lumen of the contracting branchial heart and the pericardial cavity with an injection-needle (0 · 3 mm diam.) connected tightly to a Statham pressure transducer (Pb 23 BB). This system was filled with doubly distilled de-gassed water. The signals were amplified by a bridge amplifier DMS system (6501 Burster Prâzisionstechnik) and displayed on a Kipp & Zonen recorder. This measurement system had a pressure sensitivity of 1 mm H₂O. NO leaks could be detected after the puncture of the organ. During the experiments the instruments were frequently recalibrated to avoid a drift of the base line.

2. Measurements of the total osmolality (in 27 Sepia officinalis) and the colloid osmotic (oncotic) pressures of the blood, pericardial and renal sac fluid (in three Octopus vulgaris and four Sepia officinalis) were carried out according to the method of Mangum & Johansen (1975), using a Knauer semimicro-osmometer (by means of freezing-point depression) and a Knauer membrane osmometer type 01,50, using a semipermeable bilayer cellulose-nitrate-acetate-membrane (‘extra-fine’, i.e. the membrane excluded molecules greater than mol.wt. 35000).

   The colloid osmotic measurements of the different body fluids were carried out with sea water as the reference fluid. The sensitivity and precision of the method was 1 mm H₂O.

   For each measurement, 30 μl of fluid were required to cover the semipermeable membrane. It was therefore possible to measure the blood very often, in practical 10 times/animal, while only 3–6 measurements of pericardial coelomic fluid per animal were possible.

3. Statistical analysis. Weighted means, analysis of variance, Fisher’s leastsignificant difference (LSD), and Student’s t test were carried out with a Hewlett-Packard 9815 A calculator with a 9865 A calculator plotter.

Results are given mainly for Sepia officinalis, although the organization of the significantly smaller branchial heart appendage in Octopoda shows a basically similar structure. The inner area is more or less completely covered by the pericardial coelom epithelium (folded epithelium) (Figs. 1a, b, 2 a). The folded epithelium shows, as demonstrated in detail in previous studies (Schipp et al. 1971; Schipp & Boletzky, 1975) the typical ultrastructure of transporting epithelial cells with a thick lamina basalis, basal labyrinth, brush border and a very high mitochondria content apically as well as in the basal cell region, and also apocrine secretory mechanisms (Schipp et al. 1971). The wall of the pericardial appendage, i.e. the area between the inner transportactive folded epithelium and the principally similar organized outer coelom epithelium is filled by a highly branched system of lacunae, stretching out from the lumen of the branchial heart. The continuous lamina basalis of both epithelia (folded epithelium and outer epithelium) borders the blood spaces, as do the basal parts of another cell-form, the polygonal cells, which often form rosette-like groups around the haemocyanin-containing blood lacunae (Figs, 1b, 2 a, b). The basal parts of these polygonal cells have ridge-like foot processes like the podocytes in the vertebrate nephron; hence in earlier studies these cells have been discussed as podocytes (Kümmel, 1967; Schipp et al. 1975) and are believed to be the site of the probable ultrafiltration process. As can be seen in Figs. 1b, 2 a, b, the apices of the podocytes, characterized by a sparsely microvillous border, reach the inner lumen of the organ at the distal ends of tubular infoldings where the inner folded epithelium is lacking. The cytoplasm of the podocytes is transparent and contains, in addition to dispersed mitrochondria and a few dense bodies, numerous dictyosomes and a relatively high amount of vesicles with long, primarily basal-apically orientated tubules (Fig. 1c). These tubules are basally open towards the spaces between the foot processes of the podocytes, which are separated from the blood lacunae only by the continuous lamina basalis and the diaphragmata between the foot processes of the podocytes (Fig. 1 c, 4b). The podocytes differ from the adjacent cells of the folded epithelium not only in the lack of a basal labyrinth and the sparsely microvillous border (Fig. 1 c, 4a) but also in the type of the lateral interdigitations (Fig. 1 c, 3 a, b). No tight lateral interdigitations can be found. In contrast, the intercellular space is irregularly shaped and often more or less enlarged. Moreover, desmosome-like structures and tight or gap junctions are missing. Thus the intercellular space seems to be open at the apex and only in such areas where the cells are associated more closely is it crossed by fibrous structures, obviously derived from the glycocalyx (Fig. 1 c).

The lumen joining the podocytes shows different dimensions. Often, especially in the peripheral region of the organ, it is reduced to a small, hardly recognizable slit (Fig. 2c) while in other areas it is larger, for example in the space between the branchial heart and the branchial heart appendage (Fig. 3 a, b).

The course of the blood circulation in the branchial heart complex is indicated in Fig. 1 a. The O₂-deficient blood flows during diastole from the vena cava and the anterior and posterior mantle veins into the lumen of the branchial heart, driven primarily by pulsations of the vena cava. During systole of the branchial heart and closing of the valves, the blood is pressed into the arteria branchialis, which is, at least in Sepia officinalis, separated from the branchial heart by fleshy valves (Tompsett, 1939). Considering the morphology of the organs, and based on in vivo studies on Sepia officinalis (R. Schipp, unpublished observations) it is highly likely that during the beginning of the contraction of the branchial heart systole, blood is pressed from the lumen of the branchial heart into the lacunae system of the branchial heart appendage. The back-flow of this blood takes place during the diastole of the branchial heart, and is driven by the contraction of the obliquely-striated muscle cells of the peripheral region of the appendage (Fig. 1a, 2 a), which probably possess an inherent autonomic activity and work antagonistically to the branchial heart, i.e. they contract when the branchial heart is in diastole. Accordingly, the pressure in the branchial heart lumen and the lacunae of the appendage correspond at least during systole, while the pressure during the diastole of the branchial heart is higher in the branchial heart appendage than inside the branchial heart.

The hydrostatic pressures in the branchial hearts were about the same in Sepia officinalis and Octopus vulgaris. The mean values of systole and diastole were not correlated with mantle length or body-weight of the animals (Table 1).

The beat rhythm, and relationship between systolic and diastolic pressure, varied from animal to animal (cf. Figs. 5 a-c). Thus pressure can be high and steady (Fig. 5 b), high and unsteady (Fig. 5 a), low and steady (data not shown), or low and unsteady (Fig. 5c) The pressure values of the pericardial coelom are generally significantly lower than in the branchial heart lumen and fluctuate only slightly around zero (Fig. 5d).

Measurements of the colloid-osmotic pressures yielded for the blood, with its high amount of high-molecular haemocyanin, the expected relatively high values of between 25–40 mm H₂O (Table 1). Similar values were found for both species, despite the different molecular weights of the haemocyanin. The colloid osmotic pressure of the protein-deficient pericardial fluid was significantly lower than that of the blood and exceed zero generally only by a few mm H₂O (Table 1). Subtraction of the values of the pericardial fluid from those of the blood results in the actual effective osmotic pressure difference (ΔP₀, Table 1) at the border between blood space and pericardial coelom.

This osmotic pressure must be contrasted to the hydrostatic pressure in considering a possible ultrafiltration. The means of effective filtration pressures of the branchial heart in maximal systole and minimal diastole respectively are shown in Table 1. A plus sign indicates that the mean hydrostatic pressure is higher than ΔP₀, i.e. ultrafiltration is possible. A negative sign indicates that the oncotic pressure exceeds the blood pressure. At maximal systole of the branchial heart all measured pressures of both species are higher than ΔP₀, while at minimal diastole this is only partially true. The likelihood that maximal systolic and partially diastolic pressures will exceed ΔP₀ is shown for both species in Fig. 6, where all measured diastolic and systolic values of the hydrostatic pressure are considered in frequency of their appearance.

In addition to the colloid osmotic measurements, the total osmolality of the blood and other body fluids of 27 Sepia officinalis was measured to the variably high values of the surrounding sea water of the Bassin d’Arcachon. The osmolalities of the blood, the fluids of the coelom and of the ventral and dorsal renal sac are significantly lower than that of the sea water, while there are no significant differences between the measured body-fluids (P₀ = 0·05) (Fig. 7). Furthermore, the osmolalities of the body fluids were hardly affected by the considerable fluctuations of the salinity and therefore osmolality of the surrounding sea water (Fig. 8).

In earlier morphological studies (Kümmel, 1967; Schipp et al. 1971 ; Schipp & Von Boletzky, 1975; Witmer & Martin, 1973), the branchial heart appendage of dibranchiate cephalopods is regarded as the possible site for a pressure filtration in the sense of a first step in urine formation, based on the demonstration of podocyte-like cellforms. The present cytological study shows, through the example of Sepia officinalis, that the podocytes, located at the base in blood lacunae of the peripheral wall reach with the microvillous border of their apex the inner lumen of the organ; i.e. the inner lumen surface is not completely covered by the transporting epithelium, as formerly assumed (Marceau, 1905; Kümmel, 1967; Schipp et al. 1971; Witmer & Martin, 1973).

The cytomorphological characteristics of the podocytes, i.e. the hyaline cytoplasm with numerous dictyosomes and vacuoles, the food processes, the sparse microvilli, the few mitochondria and the absence of a folded labyrinth, are in distinct contrast to those of the transporting epithelium described formerly (Schipp et al. 1971 ; Schipp & Boletzky, 1975). These findings correspond to earlier histochemical and cytochemical data; e.g. in contrast to the transporting epithelium, dehydrogenases, phosphatases and monoamineoxidase (Schipp et al. 1971) as well as Mg²⁺-and Na⁺-K⁺-ATPase (Donaubauer, 1979) and glutamic oxaloacetic transaminase (R. Schipp, unpublished observations) are hardly or not at all present in the podocytes.

The values of the colloid osmotic pressures of blood and pericardial fluid in Sepia officinalis and Octopus vulgaris, directly measured for the first time, demonstrate that, in accordance with the high protein content of the blood (Sepia: 90–100 g/l) (R. Schipp, unpublished observations), the oncotc pressures of the blood with mean values of 33·5 mm H₂O (Octopus) and 30·1 mm H₂O (Sepia) are significantly higher than those of the coelomic fluid (5·2 and 4·7 mm H₂O). The data are similar to theoretical values, calculated from the protein content (Florkin & Blum, 1934; Mangum & Johansen, 1975). For example, for Octopus blood, 42·6 mm H₂O was calculated.

The total osmolality of the body fluids in Sepia officinalis is significantly hypotonic to the sea water and indicates that the animals are able to actively counter fluctuations in the osmolalities of the body fluids that are produced by changes in the osmolality of the environment. Consequently the animals are by no means poikolosmotic and can be classified as ‘hypo-osmotic regulators’ (Potts & Parry, 1964). The probable sites of active regulations are the widespread transporting epithelia in the branchia and the various renal organs (Schipp & Boletzky, 1975) as has been shown by the demonstration of Na⁺-K⁺-ATPase (Donaubauer, 1979; unpublished 1980) and measurements of ionic composition of the body fluids (Robertson, 1953; Potts & Todd, 1965).

The measured systolic and diastolic pressures of the branchial heart show strong fluctuations depending on the size of the animal and apparently also on the stimulations induced by experimental treatments - for example, mechanical stimulations of the mantle muscle (Mislin, 1950, 1967). The systolic mean values of the Octopus vulgaris (adult) are about 46 mm H₂O (maximally 85 mm H₂O), for Sepia officinalis (adult) about 48 mm H₂O, In contrast, values obtained by different methods of other researchers are significantly higher for the genus Octopus, while for the genus Sepia corresponding values are missing. De Wilde (1956) reports 130–190 mm H₂O for Octopus vulgaris and Johanson and Huston (1962) 100–150 mm H₂O for the considerably larger Octopus dofleini, measured in the afferent branchial vessel originating from the branchial heart.

Lately Wells (1979) and Wells & Mangold (1980) have reported excellent measurements from non-anaesthetized free-moving Octopus vulgaris and have found systolil values of 60–80 mm H₂O in the afferent branchial vessel. Through the direct measurement method used by Wells, where the free-swimming animal is connected by a ‘nylon pipe’, it is questionable to what extent the mantle pressure, acting on the branchial heart, influenced the recording, thus giving rise to the relatively high values. The comparison of hydrostatic and colloidosmotic pressure between the branchial heart lumen and the pericardial coelom demonstrates that at least during the branchial heart systole, but also in a considerable period of the diastole, a pressure ultrafiltration is possible in both species, even in the absence of a complete neuronal and hormonal system (Wells, 1979; Wells & Mangold, 1980). This pressure difference is unaffected by mantle pressure and the height of the water column above the animal, since these pressures affect all compartments of the animal equally.

The circulation of the blood between the branchial heart and the branchial heart appendage is not yet completely clarified. The general morphology (the absence of valves, the lack of afferent and efferent vessel differentiation between the two organs ; Fig. 1 a) and in vivo observations indicate a kind of shuttle movement, i.e. the filling of the appendage during the systole of the branchial heart, followed by an antagonistic flowback of blood in its diastole by reflectory contractions of the obliquely striated muscle cells in the appendage. In this modus of circulation, the pressure in the appendage should be higher than that measured in branchial heart diastole and a pressure filtration could be probable also in this phase. But the formation of a functional unit of the two organs cannot be dismissed and in the event of this formation the pressure phases of both organs would be more or less identical.

Surveying the finding of this and earlier studies (Harrison & Martin, 1965 ; Schipp et al. 1971; Witmer & Martin, 1973; Donaubauer, 1979), the urine formation in branchial heart appendage can be considered as a process of ultrafiltration through the lamina basalis of podocytes, followed by a passage of the ultrafiltrate through the enlarged lateral intercellular spaces as well as intracellular vesicular passages of the podocytes and processes of reabsorption and secretion in the slit-like infoldings of the transporting epithelium. This means that already inside the branchial heart appendage, the ultrafiltrate is concentrated (e.g. a concentration of ammonia; R. Schipp, unpublished observations) ; a process which is continued in the ductus renopericardialis of Octopus as well as in the renal sac of both species (Schipp & Boletzky, 1975).

Taking into account that in the branchial heart complex breakdown products of metabolism, for instance in the form of Fe-III- and Cu-proteids, are stored (Banauch, 1970; Schipp & Hevert, 1978), it can be concluded that this organ complex performs more or less all forms of excretory activity: excrete-storage, ultrafiltration as well as processes of active reabsorption and secretion

This study was supported by the Deutsche Forschungsgemeinschaft.